用小波分解一维信号的不同子带

Question

我是研究小波的，在这个领域我是新手。我想把一个信号分解成多个band.So，我使用wavedec()把一个信号分解成5级，然后用wrcoef()重建单个band.But，问题是当我把5个频带相加时，重构信号与原始信号的差异更大。请任何人在这方面帮助我。下面是我的代码..

load sumsin; s = sumsin; 
figure;plot(s);
% Perform decomposition at level 5 of s using sym4. 
[c,l] = wavedec(s,5,'sym4');

% Reconstruct approximation at level 5, 
% from the wavelet decomposition structure [c,l].
a1= wrcoef('a',c,l,'sym4',1);
a2 = wrcoef('a',c,l,'sym4',2);
a3 = wrcoef('a',c,l,'sym4',3);
a4 = wrcoef('a',c,l,'sym4',4);
a5 = wrcoef('a',c,l,'sym4',5);


 figure; subplot(5,1,1); plot(a1); title('Approximation at level 1');
 subplot(5,1,2); plot(a2); title('Approximation at level 2');
 subplot(5,1,3); plot(a3); title('Approximation at level 3');
 subplot(5,1,4); plot(a4); title('Approximation at level 4');
 subplot(5,1,5); plot(a5); title('Approximation at level 5');
 figure;plot(a1+a2+a3+a4+a5);title('Reconstruct Original signal');

Answer 1

要重建原始信号，您需要将五个详细分量和最后一个、第五个级别的近似分量相加。

d1= wrcoef('d',c,l,'sym4',1);
d2 = wrcoef('d',c,l,'sym4',2);
d3 = wrcoef('d',c,l,'sym4',3);
d4 = wrcoef('d',c,l,'sym4',4);
d5 = wrcoef('d',c,l,'sym4',5);
a5 = wrcoef('a',c,l,'sym4',5);
s_origin=d1+d2+d3+d4+d5+a5;